Anshul Singhvi
Anshul Singhvi
FWIW I get this on main too
Both unfortunately. Yes it looks like Xarray is doing CF scaling, and the dataset is sea surface temperature, so the Xarray data range makes _way_ more sense than the Rasters/NCDatasets...
You don't for this, it's public access.
Huh! Here's a google drive link: https://drive.google.com/file/d/1Om08vXhgsmLbcGmbEFaachfBHa6Nq5pA/view?usp=sharing
Which OS/DiskArrays versions are you on? I'm on Mac M1 and DiskArrays 0.4.4
Ah, might be a DiskArrays 0.4.x thing then?
also: this plot illustrates more sharply that the nature of the values is different (some places have different signs or something pre scaling?!) between NCDatasets/Rasters and Xarray. 
Ah possibly, yeah I am on Fabian's PR branch for NCDatasets. This also produces the same issues with strange values with the same data but encoded in Zarr by the...
Couldn't get the wrapped arrays to work together because of a dimension mismatch (`Dim{:x}` v/s `X`) but `collect` saved the day: 
There's clearly some structure here that I bet is described by some CF attribute. But I don't know what it is!